Pathological endometrial function characterization in the mid-secretory phase in infertile patients

Resumen

A displaced and/or pathological window of implantation (WOI) are the main potential causes described for endometrial recurrent implantation failure (RIF) in in vitro fertilization (IVF) patients. The displaced WOI is well characterized and can be diagnosed using transcriptomic tests, although clinical benefits are controversial. However, the molecular mechanisms underlying the pathological WOI remain poorly understood and require further research to characterize its molecular heterogeneity and develop alternative diagnostic tools. This thesis aimed to identify and characterize the pathological WOI in IVF patients undergoing hormone replacement therapy (HRT). Endometrial biopsies were collected during the mid-secretory phase from patients recommended for endometrial evaluation in a multicentric prospective study. The whole transcriptome from samples was analysed using RNA-Seq and gene expression data were pre-processed by removing the endometrial progression effect. Patients (n = 131) were clinically classified as RIF (¿ 3 implantation failures, n = 32) or controls (< 3 implantation failures, n = 99), and subsequently divided into training (n = 105) and test (n = 26) sets. A supervised, probabilistic artificial intelligence model was implemented to stratify the patients. Pregnancy (PR), ongoing pregnancy (OPR), biochemical and clinical miscarriage rates (BMR; CMR) were calculated with the first embryo transfer after biopsy collection. Cumulative PR was also calculated. Differential expression and functional analyses were performed and qPCR was used to validate selected genes of interest. Independent of displacements, the pathological WOI signature we identified in our HRT population paved the way for a new transcriptomic taxonomy with a gradient of prognosis. Based on the probability of pathology given by the model, patients predicted as RIF were stratified into p1 (n = 24) and p2 (n = 14), and patients predicted as controls, into c2 (n = 32) and c1 (n = 61). Compared to control profiles, pathological profiles had lower pregnancy rates (71-78% vs. 29-57% PR and 76-91% vs. 50-57% OPR, respectively) and higher miscarriage rates (0-8% vs. 12-43% BMR and 9-17% vs. 0-43% CMR). The c1 profile had the highest OPR (91%), while c2, which slightly resembles the pathological profiles, was associated with excessive sensory perception and hormonal response. In contrast, p1 was associated with the lowest PR (29%) and the highest BMR (43%), while p2 was related to the highest CMR (43%). The cumulative PR was lower in pathological profiles, showing an increasing trend from p1 to c1 (38%-93%). Accordingly, p1 was associated with an excessive immune response against the embryo in early pregnancy stages, while p2 was initially more immune-tolerant but showed embryo rejection in late stages due to the lack of metabolic response. This new taxonomy and our molecular findings set the foundation for a new generation of endometrial-factor evaluation tools towards precision medicine.